1431
Biophysical controls on surface fuel litterfall and
decomposition in the northern Rocky Mountains,
USA
Robert E. Keane
Abstract: Litterfall and decomposition rates of the organic matter that comprise forest fuels are important to fire management, because they define fuel treatment longevity and provide parameters to design, test, and validate ecosystem models.
This study explores the environmental factors that control litterfall and decomposition in the context of fuel management
for several major forest types in the northern Rocky Mountains (Idaho and Montana), USA. Litterfall was measured for
more than 10 years using semiannual collections of six fine fuel components (fallen foliage, twigs, branches, large
branches, logs, and all other canopy material) collected from a network of 1 m2 litterfall traps installed at 28 plots across
seven sites. Decomposition of foliage, twigs, branches, and large branches were measured using litter bags installed on
five of the seven sites. Measured litterfall and decomposition rates were correlated with major environmental and vegetation variables using regression analysis. Annual foliage litterfall rates ranged from 0.057 kgm–2year–1 for dry Pinus ponderosa Dougl. ex Laws. stands to 0.144 kgm–2year–1 on mesic Thuja plicata Donn ex D. Don stands and were correlated
with the vegetation characteristics of leaf area index, basal area, and tree height (r > 0.5), whereas decomposition rates
were correlated with the environmental gradients of temperature and relative humidity (r > 0.4).
Résumé : Les taux de chute de litière et de décomposition de la matière organique qui constituent les combustibles forestiers sont importants pour la gestion du feu parce qu’ils déterminent la longévité du traitement des combustibles et fournissent les paramètres pour mettre au point, tester et valider les modèles d’écosystème. Cette étude explore les facteurs
environnementaux qui régissent la chute de litière et la décomposition dans le contexte de la gestion des combustibles de
plusieurs types forestiers importants de la partie nord des montagnes Rocheuses en Idaho et au Montana, aux États-Unis.
La chute de litière a été mesurée pendant plus de 10 ans en effectuant une collecte semi-annuelle des composantes des
combustibles légers (feuilles, rameaux, branches, grosses branches, billes et tous les autres matériaux de la canopée tombés
au sol) grâce à un réseau de trappes à litière de 1 m2 installées dans 28 parcelles réparties dans sept stations. La décomposition des feuilles, des rameaux, des branches et des grosses branches a été mesurée à l’aide de sacs à litière installés dans
cinq des sept stations. Les taux de chute de litière et de décomposition ont été corrélés aux principales variables environnementales et à celles de la végétation à l’aide de l’analyse de régression. Le taux annuel de chute de litière sous forme
de feuillage variait de 0,057 kgm–2an–1 dans les peuplements xériques de Pinus ponderosa Dougl. ex Laws. à 0,144
kgm–2an–1 dans les peuplements mésiques de Thuja plicata Donn ex D. Don et il était corrélé (r > 0,5) aux caractéristiques de la végétation par l’indice de surface foliaire, la surface terrière et la hauteur des arbres tandis que le taux de décomposition était corrélé (r > 0,4) aux gradients environnementaux de température et d’humidité relative.
[Traduit par la Rédaction]
Introduction
The successful suppression of many wildland fires in the
western United States and Canadian ecosystems over the
last 70 years has resulted in increased accumulations of surface fuels that have increased the potential for severe fires
(Ferry et al. 1995). Many government agencies are advocating extensive fuel treatments and ecosystem restoration activities to reduce the severity of these intense wildfires that
could potentially damage ecosystems, destroy property, and
take human life (Laverty and Williams 2000; GAO 2002).
Knowledge of fuel litterfall deposition and decomposition
rates could help managers prioritize, design, and implement
Received 20 July 2007. Accepted 9 January 2008. Published on
the NRC Research Press Web site at cjfr.nrc.ca on 3 May 2008.
R.E. Keane. USDA Forest Service, Rocky Mountain Research
Station, Missoula Fire Sciences Laboratory, 5775 Highway 10
West, Missoula, MT 59808, USA (e-mail: rkeane@fs.fed.us).
Can. J. For. Res. 38: 1431–1445 (2008)
more effective fuel-treatment programs. However, these
rates remain relatively unknown for many forest ecosystems.
This is especially true for down dead woody fine fuels, because most studies determine only leaf litter or large log fuel
decay and accretion rates (Harmon et al. 1986).
Quantification of surface fuel dynamics across managed
landscapes is important to fire managers and researchers for
many reasons. Rates of fuel buildup and decomposition can
be used to define temporal (how long will a treatment last)
and spatial (what areas are best to effectively treat) limits to
fire hazard reduction treatments (Fernandes and Botelho
2003). These rates could also help determine how fast
treated landscapes would reach undesirable fuel loadings to
warrant another treatment. Fuel and fire modeling efforts
need litterfall and decomposition rates to realistically simulate fuel dynamics across landscapes to compare alternative
fuel treatment strategies (Keane et al. 1996). The rates can
also be used as validation of simulated ecological processes
in existing and future ecosystem process models (Pastor and
doi:10.1139/X08-003
#
2008 NRC Canada
1432
Post 1985; Botkin 1993). Carbon fluxes can also be approximated from measured fuel dynamics rates to determine the
contribution of the fuelbed to atmospheric carbon sources
and sinks (Thornton et al. 2002).
This study quantified the rates of forest litterfall and decomposition for a number of forest types across the northern
Rocky Mountains (Idaho and Montana, USA) to estimate
fuels parameters for use in complex landscape models of
fire and vegetation dynamics (Keane et al. 1996; White et
al. 1998). Because it is impossible to measure fuel deposition and decomposition for all stand types in all northern
Rocky Mountain ecosystems, the estimated rates were then
correlated to a number of biophysical and vegetation variables that were either directly estimated at the field plots or
simulated with ecosystem process models so that these rates
of fuel dynamics could be extrapolated across all areas of
northern Rocky Mountains landscapes.
Six surface fuel components are recognized in this study.
Freshly fallen leaves and needles from trees, shrubs, and
herbs are considered foliage, whereas all other nonwoody
material, such as fallen cones, bark scales, lichen, and bud
scales, are lumped into a category called other canopy fuels.
The woody material is sorted into four diameter classes using the size ranges required by the fire behavior and effects
models (Fosberg 1970; Rothermel 1972; Reinhardt and
Keane 1998). The smallest size class, called twigs, defines
1 h timelag (time it takes to dry or wet the fuel particle to
67% of its equilibrium moisture content) fuels with diameters <6 mm. Branches with diameters between 6 and
25 mm are 10 h timelag fuels, and large branches with diameters ranging from 25 to 75 mm are 100 h timelag fuels.
The logs, downed woody fuels >75 mm in diameter, define
the 1000 h timelag fuel component (Hagan and Grove
1999), which does not include snags or stumps. In this paper, the fuels described above are considered ‘‘litter’’ for
simplicity; devices used to collect fuels are referred to as litter traps. Fuel accumulation is considered litterfall minus decomposition. Duff decomposition rates, along with tree,
shrub, and herbaceous growth rates, were not measured in
this study, because there are abundant other efforts in these
areas. This study is primarily concerned with the dynamics
of the foliage, fine woody, and log surface fuel components.
Most litterfall studies have only measured the rate of foliage or log deposition (Harmon et al. 1986; Vogt et al. 1986)
(Table 1). Small woody debris additions to the forest floor,
such as twigs and branches, are rarely reported, even though
they may be the most important to fuels management and
fire behavior prediction because they contribute to fire
spread (Rothermel 1972). Deposition rates for logs (Table 1)
are usually measured from historical tree mortality and snag
fall rates over time, which assumes tree fall is the only input
to log accumulation. However, large branches and tree tops
also contribute to log inputs to the forest floor in some ecosystems (Harmon et al. 1986).
Studies of decomposition rates are usually for only the foliage and large log material, especially in the western United
States (Table 1), much like litterfall studies. The exceptions
are Edmonds (1987) and Taylor et al. (1991), who measured
decay of twigs, branches, and cones, and Carlton and Pickford (1982) and Christiansen and Pickford (1991), who estimated small wood losses by sampling different aged timber
Can. J. For. Res. Vol. 38, 2008
slash. These studies usually use the parameter k in an exponential decay curve to describe rates of decay (Olson 1963;
Robertson and Paul 2000).
Several studies have attempted to relate litterfall and decomposition rates to environmental variables. Mackensen et
al. (2003) successfully correlated values of the decomposition rate k to rainfall, temperature, and altitude using decomposition studies conducted around the world, whereas
Johannsson (1994) estimated decomposition along a latitudinal gradient. Log loadings have been correlated with topographic gradients of slope, aspect, and landform (Jenkins et
al. 2004; Sanchez-Flores and Yool 2004; Webster and Jenkins 2005) and fire regime (White et al. 2004; Brais et al.
2005). Huebschmann et al. (1999) correlated needle fall to
stand, site, and weather characteristics and found that spring
temperatures were the best predictors. Meentemeyer (1978)
developed a global model of litter production using climate
factors, such as evapotranspiration.
It is difficult for fire managers to determine litterfall and
decomposition rates at the stand level and almost intractable
at the landscape level, because it requires extensive networks
of collection devices that must be frequently monitored over
long time periods (5–10 years) to accurately estimate annual
fluxes. The density and spacing of the collection devices are
highly dependent on the type of fuel collected. Coarse
woody fuels usually require installing larger traps across
larger areas and are monitored for longer time periods,
whereas fine fuels would require smaller traps but frequent
visitation to minimize decomposition losses. An alternative
would be to predict fuel dynamics from those environmental
variables that control deposition and decomposition. The
objectives of this study were to (i) measure rates of litterfall
and decomposition for important fuel components, (ii) determine important relationships of biophysical variables to the
measured rates, and (iii) create empirical predictive equations that could be used to extrapolate stand-level fuel dynamics estimates across entire landscapes to support spatial
fuels modeling efforts and landscape planning.
Materials and methods
Study sites
This study began as an extension of two previous studies
that explored the use of ecosystem modeling and gradient
analysis to create digital maps of current and future landscape characteristics. In 1993, a set of litter traps were installed on two sites in western Montana to parameterize and
validate two ecosystem models: BIOME-BGC (White et al.
2000) and Fire-BGC (Keane et al. 1996) (Table 2, sites CO
and SB). Then, in 1995, an extensive field sampling project
was initiated to explore the use of measured and simulated
environmental gradients to map ecosystem characteristics,
such as fuels, across landscapes (Keane et al. 2002; Rollins
et al. 2004). To validate the models used in the two studies,
the number of sites was expanded from two to six by establishing four new sites along elevational and aspect gradients
within the larger northern Rockies study area (Fig. 1). In
1997, one more site was included to represent the ubiquitous
lodgepole pine (Pinus contorta Dougl. ex Loud.) ecosystem
that occurs east of the Continental Divide (Table 2, site TF;
Fig. 1). Other forest types represented by these sites include
#
2008 NRC Canada
Keane
Table 1. Litterfall and decomposition rates for foliage and woody material measured in various western US ecosystems.
Litterfall rate
(kgm–2year–1)
Decay constant
k (year–1)
Province or state
Reference(s)
0.03
0.29
0.05
0.05, 0.14, 0.08–0.18
Arizona
California, Arizona
Avery et al. 1976; Klemmedson 1992
Bray and Gorham 1964; Yavitt and Fahey 1982; Stohlgren 1988;
Klemmedson et al. 1990; Hart et al. 1992
0.70, 0.45, 0.04,
0.15–0.45, 0.28
0.006–0.050
Oregon, Washington
Twigs and branches
—
Washington
Foliage
0.50, 0.17–0.33,
0.114–0.177
0.007–0.129, 0.06–0.14,
0.06, 0.005–0.05
0.005–0.010, 0.44, 0.27,
0.41–0.56, 0.178–0.284
Wright and Lauterback 1958; Grier and Logan 1977; Gottfried 1978;
Sollins 1982; Harmon et al. 1986; Spies et al. 1988; Edmonds and
Eglitis 1989; Harmon and Hua 1991
Fogel and Cromack 1977; Edmonds et al. 1986; Edmonds 1987; Edmonds
and Eglitis 1989; Christiansen and Pickford 1991; Maguire 1994
Dimock 1958; Turner and Long 1975; Fogel and Cromack 1977; Edmonds
1979; Graham 1982; Means et al. 1985; Sollins et al. 1987; Edmonds
1991; Harmon and Hua 1991; Trofymow et al. 1991; Prescott et al. 2000
Ecosystem and
fuel component
Pinus ponderosa
Logs
Foliage
Pseudotsuga menziesii
Logs
Pinus contorta
Logs
Twigs
Foliage
Tsuga heterophylla
Logs
Twigs and branches
Foliage
Abies lasiocarpa – Picea spp.
Logs
Foliage
0.02
Oregon, British
Columbia,
Washington
Colorado, Alabama
—
0.362
0.027, 0.082, 0.0016–
0.0027, 0.115, 0.015
0.055
0.115, 0.14, 0.09–0.11
—
—
—
0.016–0.018
0.08–0.24
0.3–0.5
Oregon
Washington
British Columbia
Graham 1982
Edmonds 1987
Keenan et al. 1996
—
0.2–0.23
0.001–0.0015
0.09–0.17
Colorado
Alabama
Kueppers et al. 2004
Taylor et al. 1991; Laiho and Prescott 1999; Prescott et al. 2003
Alabama
Alabama, Wyoming
Alexander 1954; Pearson et al. 1987; Taylor et al. 1991; Busse 1994;
Laiho and Prescott 1999; Kueppers et al. 2004
Taylor et al. 1991 ; Prescott et al. 1993
Yavitt and Fahey 1982; Taylor et al. 1991; Berg and Ekbohm 1993; Stump
and Binkley 1993; Laiho and Prescott 1999
Note: Twigs and branches are assumed to be less than 8 cm (3 in.) in diameter.
#
1433
2008 NRC Canada
1434
Can. J. For. Res. Vol. 38, 2008
Table 2. General description of the study sites and plots included in this study.
Study site and plot
Coram (CO)
1
2
3
4
Snowbowl (SB)
1
2
3
4
Red Mountain (RM)
1
2
3
4
Spar Lake (SL)
1
2
3
4
Red River (RR)
1
2
3
4
Keating Ridge (KR)
1
2
3
4
Tenderfoot (TF)
1
2
3
4
Cover
typea
Habitat
typea
Elevation
(m)
Aspectb
Collection
years
Basal area
(m2ha–1)c
DF/WL
WC/WH
SF
WP
GF/CU
WH/CU
SF/MF
SF/XT
1185
1184
1937
1915
SW
NW
NE
SW
1993–2005
1993–2005
1993–2005
1993–2005
29.87
50.44
10.58
34.34
PP
DF
LP
SF/WP
DF/VS
DF/PM
SF/XT
SF/MF
1680
1596
1972
2073
NW
S
SW
E
1995–2005
1995–2005
1995–2005
1995–2005
PP
WC/WH
WP
SF
DF/CR
WH/CU
SF/XT
SF/MF
943
942
1988
1529
E
E
SE
NW
WC
DF
WC
WL
WH/CU
GF/XT
WH/CU
WH/CU
1090
1124
1260
1600
PP
GF/DF
LP
LP
DF/LB
GF/LB
SF/XT
SF/XT
GF
PP
LP
SF
LP
LP/SF
LP
LP
Tree density
(stemsha–1)c
Fuel loading
(kgm–2)d
LAI
(m2m–2)e
296.4
741.0
222.3
938.6
27.26
1.84
8.43
18.45
1.75
2.24
0.63
3.10
31.28
36.57
30.11
32.76
864.5
666.9
988.0
568.1
1.02
1.37
2.61
2.19
2.85
2.77
1.74
3.17
1995–2005
1995–2005
1995–2005
1995–2005
34.96
55.68
19.00
31.31
197.6
395.2
395.2
395.2
3.10
28.12
2.22
8.42
4.01
3.38
1.81
2.47
SE
S
S
SE
1995–2005
1995–2005
1995–2005
1995–2005
64.85
48.48
52.71
68.22
1284.4
419.9
988.0
617.5
9.67
9.33
6.02
19.87
7.90
6.58
6.10
7.02
1425
1407
1988
1979
N
SW
W
E
1995–2005
1995–2005
1995–2005
1995–2005
37.41
35.42
28.65
32.32
345.8
172.9
543.4
889.2
19.94
4.82
6.27
6.98
4.40
2.69
2.21
2.69
GF/LB
PP/SA
SF/XT
SF/XT
1041
1340
2004
2078
E
W
W
E
1995–2005
1995–2005
1995–2005
1995–2005
46.53
47.35
51.31
70.72
518.7
345.8
1630.2
1654.9
20.09
10.12
2.35
5.20
8.39
3.01
4.41
6.59
SF/VS
SF/VS
SF/VS
SF/VS
2302
2299
2143
2158
F
F
F
F
1997–2005
1997–2005
1997–2005
1997–2005
53.75
44.26
25.95
38.23
1309.1
839.8
716.3
1284.4
2.76
0.47
0.71
0.77
4.24
3.31
2.23
3.38
a
Cover type and habitat type species are as follows: trees are PP, ponderosa pine (Pinus ponderosa var. ponderosae); DF, Douglas-fir (Pseudotsuga
mensezii); WL, western larch (Larix occidentalis Nutt.); WC, western red cedar (Thuja plicata), WH, western hemlock (Tsuga heterophylla); LP,
lodgepole pine (Pinus contorta var. contorta); WP, whitebark pine (Pinus albicaulis); SF, subalpine fir (Abies lasiocarpa); GF, grand fir (Abies grandis
(Dougl.) Lindl.); and undergrowth species are CR, Calamagrostis rubescens Buckley; CU, Clintonia uniflora (Menzies ex Schult Schult. f.) Kunth; LB,
Linnaea borealis L.; MF, Menziesia ferruginea Sm.; PH, Physocarpus malvaceus (Greene) Kuntze; VS, Vaccinium scoparium Leiberg ex Coville; XT,
Xerophyllum tenax (Pursh) Nutt. Cover types are based on plurality of basal area and habitat types are from Pfister et al. (1977).
b
Aspect codes are as follows: N, north; S, south; E, east; W, west; F, flat.
c
Only overstory trees (>10 cm DBH) were used to compute basal area and density.
d
Fuel loading only includes downed dead woody fuels summed across all four size classes.
e
LAI, projected leaf area index.
pure or mixed stands of ponderosa pine (Pinus ponderosa
Dougl. ex Laws.), Douglas-fir (Pseudotsuga menziesii
(Mirb.) Franco), western red cedar (Thuja plicata Donn ex
D. Don), subalpine fir (Abies lasiocarpa (Hook.) Nutt.), and
whitebark pine (Pinus albicaulis Engelm.) (Table 2).
Each study site consisted of four 0.04 ha circular plots established along the major topographic gradients of elevation
and aspect. Plots were established in readily accessible areas
at low and high elevations and on north and south aspects to
adequately describe the diversity of important direct environmental gradients, such as productivity, moisture, and
temperature (Keane et al. 2002). At each plot, a wide variety
of topographic, vegetation, and ecosystem variables were
measured or estimated on plots using sampling protocols in
the ECODATA sampling package (Hann et al. 1988). The
entire list of sampled attributes is provided in Keane et al.
(2002), but the most important among them are an inventory
of all trees in the plot to compute basal area, leaf area, and
stand density and a network of 30 m fuel transects (Brown
1970) to estimate fuel loadings for the six fuel components
used in this study. Details of all methods used in this study,
especially litterfall collection and decomposition measurement discussed next, can be referenced in Keane (2008).
Measuring litterfall
Within each plot, seven to nine litter traps were placed on
the forest floor in the pattern shown in Fig. 2 to collect
fallen biomass. Nine litter traps were established at the two
sites installed in 1993 (Table 1, sites CO and SB), but a subsequent analysis of variance of fallen foliage and woody material in 1995 showed that only seven traps were needed to
adequately sample litterfall. Litter traps were constructed by
#
2008 NRC Canada
Keane
Fig. 1. Geographic locations of the seven sites included in this
study. Four litter collection plots were established at each of the
seven sample sites at high and low elevations and on north and
south aspects.
creating a 1 m 1 m frame with 2 cm 14 cm boards and
then tacking a coarse grid hardware cloth on the bottom of
the frame to allow water to rapidly drain from the trap and
minimize losses from accumulated material due to decomposition and wind. A plastic screen (mesh size 0.7 mm) was
secured on top of the hardware cloth at the bottom of the
trap to block fine material from falling through the coarse
hardware grid.
Each plot was visited once a month during the snow-free
periods of the year, and all materials in each trap were
placed into heavy paper bags. Woody fuel particles that lay
partially out of the trap were sawed directly at the trap border as defined by the inside dimension of the trap boards.
An estimate of projected leaf area index, (LAI, m2m–2) was
taken with a LI-COR LAI-2000 (LI-COR Inc., Lincoln,
Neb.) during each plot visit to document any major changes
in the forest canopy using the Nackaerts et al. (2000) methods. The monthly visits were designed to minimize mass
losses due to decomposition as the fallen material sat in the
traps; however, because little decomposition was observed
during the hot, dry months of summer, it was determined
that the most critical times for sampling were directly after
snowmelt and just before the first autumnal snows. Therefore, starting in 2002, all the plots were visited at these two
times each year.
The collected materials were transported to the laboratory,
and the labeled bags were placed in an oven set at 80 8C for
2 or 3 days. The dried litter was then placed in cake tins and
sorted by hand into the six fuel components (foliage, twigs,
branches, large branches, logs, and other canopy material)
and then weighted to the nearest 0.01 g. A small sample of
the dried material was set aside for the decomposition measurements.
Estimating decomposition
Litterbags were used to estimate the rate of decay for four
fuel components: freshly fallen foliage, twigs, branches, and
1435
Fig. 2. Position of the litter traps within a plot was in a cross-like
pattern with the Coram and Snowbowl sites (Table 2, sites CO and
SB) having nine litter traps and the remaining sites having seven
traps per plot (NW and SE were missing). Circles are at 11.6 m and
5 m in radius from plot center and numbers represent azimuths.
Acronyms reference compass directions: N, north; NW, northwest,;NE, northeast; E, east; S, south; SE, southeast; SW, southwest; W, west; PC, plot center.
large branches (Prescott et al. 2000). The bags were made
by sewing together with UV resistant thread a fiberglass
screen with a pore size of about 2 mm for the top with a rumen bag or pool cover material with a pore size of
0.055 mm for the bottom (Keane 2008). Bags for foliage
were roughly 170 mm 170 mm, and bags for the woody
fuels were roughly 170 mm 130 mm (0.0221 m2). Approximately 100–150 g of freshly fallen material taken from
the litter traps (see previous section) was placed into each
bag, and then the bag was sewn closed. The bags with
wood were dried at 50 8C for 3 days, and the foliage bags
were air-dried (the low oven temperature was used to minimize chemical changes) and then weighed to the nearest
0.01 g. Log, duff, and other canopy material decomposition
rates were not estimated because of limited time, lack of appropriate equipment, and logistical considerations.
Three sets of three bags each (nine bags total) for the
three fine woody fuel components (twigs, branches, and
large branches) and three sets of six bags each (18 bags) for
the foliage material were installed at each plot. A set from
each of the four fuel components (foliage, twigs, branches,
and large branches) was placed near plot center, another set
of four was placed at about 7 m northwest of the plot center,
and the third about 7 m southeast of the plot center. The litter bags were laid on top of the litter layer in late autumn
and secured using a wire that was sewn through each bag
and attached to a large 20 cm spike driven into ground to a
depth of 19 cm to prevent downslope movement and minimize ungulate damage. Decomposition was estimated over
3 years with one foliage bag taken from each wire set every
#
2008 NRC Canada
1436
Can. J. For. Res. Vol. 38, 2008
Table 3. Sampled, summarized, and simulated variables used to correlate with the litterfall and decomposition rates
to develop predictive models.
Variable name
Topographic features
ASPECT
ELEV
SLOPE
Vegetation characteristics
AGE
AVEDBH
AVEHT
BAREA
CLAY
DOMDBH
DOMHT
FUELLOAD
LAI
MAXAGE
SAND
SAPAREA
SAPPH
SILT
SNAGDBH
SNAGBA
SOILDEPTH
SPH
TPH
Environmental gradients
NPP
OUTFL
PPT
SRAD
TDAY
TMAX
TMIN
VPD
Description
Source
Units
Direction of exposure
Elevation
Slope of plot
Field data
Field data
Field data
degrees
m
%
Mean tree age
Mean diameter at base height
Mean tree height
Overstory basal area
Percent clay in soil
Dominant diameter at base height
Dominant tree height
Fuel loading
Leaf area index
Maximum tree age
Percent sand in soil
Sapling basal area
Sapplings per hectare
Percent silt in soil
Snag diameter at base height
Snag basal area
Soil depth for 90% rooting zone
Snags per hectare
Trees per hectare
Field data tree list
Field data tree list
Field data tree list
Field data tree list
Direct measurement
Field data tree list
Field data tree list
Field data tree list
Field data tree list
Field data tree list
Direct measurement
Field data tree list
Field data tree list
Direct measurement
Field data tree list
Field data tree list
Direct measurement
Field data tree list
Field data tree list
years
cm
m
m2ha–1
%
cm
m
kgm–2
m2m–2
years
%
m2ha–1
treesha–1
%
cm
m2ha–1
m
ha
m2ha–1
Net primary production
Soil water outflow
Mean annual precipitation
Mean annual daily solar radiation
Mean annual day temperature
Mean annual maximum temperature
Mean annual minimum temperature
Mean annual vapor pressure deficit
BIOME-BGC model
BIOME-BGC model
DAYMET weather
DAYMET weather
DAYMET weather
DAYMET weather
DAYMET weather
DAYMET weather
kg Cm–2
kg H2Om–2
cm
kJm–2day–1
8C
8C
8C
mbar
Note: The term ‘‘tree list’’ signifies that the variable was summarized from individual tree data collected using methods described in Keane (2008). The DAYMET and BIOME-BGC models are described in the text.
6 months, and one bag from each of the three woody components was taken every 12 months. The retrieved bags
were cut from the wire and any material that had fallen
onto the bag or that was attached to the bottom of the bag
was scraped off using a knife. The litterbags were placed in
paper bags and brought back to the laboratory to be dried at
50 8C for 3 days and weighed.
Quantifying environmental variables
Biophysical and environmental variables were quantified
for each plot using a number of techniques detailed in Keane
et al. (2002) and Rollins et al. (2004). Site and vegetation
descriptions were taken from the sampled data or from summaries of the data (Keane 2008). These variables included
general stand measurements (e.g., tree cover, bare soil cover,
and shrub cover) and site observations (e.g., elevation, aspect, and slope) at the plot level. Computed values summarized from sampled data included basal area and tree density
from individual tree measurements of height, diameter, and
species (see Table 3 for the most important variables).
Some of the sampled plot data and corresponding summa-
ries were used as inputs to complex climate and ecosystem
models (Rollins et al. 2004). Climate variables were computed from daily weather extrapolated to the plot from a network of base weather stations using the DAYMET model
(Thornton et al. 1997). Annual means of the simulated climate variables were computed over the entire 18 year length
of record of the base weather stations. Additional plot data
coupled with the simulated daily DAYMET weather data
were used as inputs to the BIOME-BGC model to simulate
a number of ecosystem process variables, such as annual net
primary productivity, evapotranspiration, and respiration
(Running and Hunt 1993; Thornton et al. 2002). These process variables were estimated by executing the BIOME-BGC
model until equilibrium (usually about 1500 years) and then
averaging the annual estimates of the process variables over
the same weather record time period used for the climate
variables (18 years). In all, there were over 50 simulated
variables available to use in the environmental analysis, but
past studies found that many of these variables are highly
correlated with each other and some gradients have little
value in predicting ecosystem characteristics and processes
#
2008 NRC Canada
Keane
(Keane et al. 2002; Rollins et al. 2004). The final list of
computed and simulated variables used in this study is
shown in Table 3.
Analyzing collected and simulated data
Annual litterfall rates (kgm–2year–1) were computed by
dividing the total amount of accumulated material across
the entire sampling record by the number of days in that record to get a daily rate, and that daily rate was multiplied by
365 to estimate an annual rate (Keane 2008). Two estimates
of decomposition were calculated: (i) k was estimated by parameterizing the exponential decay function below using regression analysis, and (ii) a mass loss rate (% biomass
lossyear–1) was estimated from differences in bag masses
over the 3 year period. The analysis to determine the decomposition parameter k in the Olson (1963) equation was
performed in SPLUS (Venables and Ripley 1999) using a
linear mixed effects model whose form is as follows:
xij
½1
ln
¼ ðk þ bi Þtj þ "ij
xi0
where xij is the mass of the ith trap at time j (tj) and xi0 is
the initial mass of the ith trap; bi is the random effect of
trap i representing the deviation of the slope from the fixed
effect for trap i; and "ij is the random error, which is assumed to be independently distributed with a normal distribution.
The computed annual litterfall and decomposition rates
were correlated with the measured and simulated vegetation,
topographic, climate, and ecosystem variables in Table 3 to
determine possible empirical predictive relationships. We
used a Pearson’s correlation coefficient (r) threshold of 0.40
to identify critical relationships. Multivariate least-squares
regression analyses were used to create predictive models of
fuel dynamics from the biophysical gradients for land management applications. A stepwise procedure was used based
on the Akaike’s information criterion (AIC) to determine the
‘‘best’’ regression equation that contained at most, two predictive variables (only two variables were used because of
low degrees of freedom and variables were deleted when
p < 0.05). The data were transformed, if needed, to meet
the assumptions of linear regression. All linear regressions
were performed in SPLUS, and residual plots were examined to check the validity of the regression model (Venables
and Ripley 1999).
Results
Litterfall and decomposition rates
Foliage had the highest litterfall rates of all five six fuel
components (0.068–0.23 kgm–2year–1) across all sites, but
interestingly, the deposition rates for the fine woody fuel
and other canopy material components were similar across
most of the sites (0.01–0.05 kgm–2year–1; Fig. 3a; see also
Keane 2008). The highest litterfall rates occurred on plots
with northern exposure, high basal area, high LAI, and
cover types composed of shade-tolerant species (Fig. 3b).
Log fall was recorded in only 47% (15 of 28) of the plots
across all traps for the entire ‡10 year recording period;
however, 90% of the plots experienced large branch fall,
and all plots recorded foliage, twig, branchwood, and other
1437
Fig. 3. Distribution of litterfall rates by surface fuel component for
all fuel collected in all litter traps (a) across all plots and (b) for
each plot (site abbrevaitions and plot numbers are given in Table 1). In the boxplots, the lower boundary of each box is the first
quartile (25th percentile), the upper boundary is the third quartile
(75th percentile), and the line within the box represents the median
of the distribution. The upper and lower error bars are the 10th and
90th percentile, and the solid circles below and above the error bars
represent outlying values.
canopy material depositions. Foliage, twigs, and other canopy material were recorded in all traps for nearly all of the
visits (99.8%). Annual variation of litterfall rates was low
(approximately 10% of annual mean) for fine fuel components but tended to increase with increasing fuel size, probably because large fuels were rarely found in the traps.
The slowest decomposition rates were measured in the
low-elevation, southern aspect forests, especially those with
high LAI (Keane 2008). Decay rates were the highest in the
most productive sites, namely those sites on low-elevation
north aspects or high-elevation south aspects (Pfister et al.
1977). Decay rates were higher and more variable for foliage (k = 0.085–0.283 year–1) than for woody fuel (k =
0.045–0.125 year–1) (Figs. 4a and 4b). As expected, large
woody fuels had lower decay and mass loss rates than the
smaller size classes, but many sites had the same decay rates
across all woody size classes. The most productive sites
(SB-2, KR-2, and TF-2) had woody decay rates that were
equivalent to foliage decay rates. The low variability of
#
2008 NRC Canada
1438
Fig. 4. Distribution of (a) decay rates (k values) and (b) mass loss
rates (proportion loss) across all 28 plots in this study by the four
fuel components. Decomposition was measured over a 3 year period, but it was not measured for logs and other canopy material.
Boxplot parameters are as defined in Fig. 3.
Can. J. For. Res. Vol. 38, 2008
environmental variables (Fig. 5; Table 4). Surprisingly, net
primary production (NPP), an important ecosystem process
variable often associated with litterfall, had little relation to
litterfall rates, possibly because of inaccurate simulation parameters and the coarse scale (1 km2) of the integrated
weather variables. High variability of large branch and log
rates resulted in low correlation coefficients (r < 0.20).
Decomposition rates (k values) were more correlated to
environmental conditions than vegetation conditions (Fig. 5;
Table 5). Foliage and fine woody fuel decay rates were significantly correlated with maximum and mean temperatures
(r > 0.49) and vapor pressure deficit (r > 0.47) (Fig. 5), but
correlations for the coarse woody fuels are not statistically
significant for any variable (Table 5). Decay rates appeared
to be negatively correlated with tree diameter measurements
(Table 3, DOMDBH, AVEDBH, and SNAGDBH) and LAI,
but the correlation was only significant for twig woody fuel
decomposition (p < 0.05).
In the regression analysis, vegetation-based variables were
found to be the best predictors of both litterfall rates and decomposition rates (Table 6). Fine fuel litterfall, especially
foliage, had the strongest relationship to vegetation variables
(R2 = 0.50–0.78), but these same variables had poor predictive ability for decomposition (R2 = 0.16 for foliage and
branches). I expected environmental variables to have more
predictive ability for decomposition (Table 6), but vegetation variables proved to have the highest value for decay
prediction. As in the correlation analysis, it appears that
fine fuel (foliage and twigs) litterfall and decomposition
rates are correlated to canopy-related stand variables (LAI,
BAREA, and DBH), whereas the larger fuels are related to
environmental variables (SRAD) or surface fuel loadings
(FUELLOAD). Basal area seems to be a variable that can
be used for all fuel components.
Discussion
woody fuel decay would suggest low correlation with site
environment.
Biophysical relationships
As expected, the vegetation variables were better predictors of litterfall rates than the environmental variables
(Table 4), even though most correlation coefficients were
low (r < 0.85). Vegetation characteristics associated with
the amount of canopy material (LAI, basal area, sapwood
basal area, and tree height) had the highest correlations with
litterfall rates, especially the foliage fuel component (Fig. 5).
Tree density, mean diameter, and age variables were poor
predictors of litterfall, but tree height seemed to be useful
for predicting fine fuel deposition (Fig. 5; Table 4). Interestingly, the initial fuel loadings (FUELLOAD, Table 4) measured on the plot when the traps were installed (Table 2) had
little predictive value for either litterfall or decomposition.
LAI appeared to be the best variable to predict litterfall for
all fuel components (r > 0.42, p < 0.05; Fig. 5), but snag
basal area also appears important (r > 0.46 for all fuel components). Only percent clay soil (CLAY), minimum temperature (TMIN), and solar radiation (SRAD) had significant
correlation coefficients with fine fuel litterfall rates for the
Litterfall
Litterfall rates in this study are slightly lower than those
in other studies (compare Table 1 with Fig. 3), probably because the northern Rocky Mountain forests are less productive than the Pacific Northwest forests in Table 1 (Harmon et
al. 1986). The low elevation moist sites of this study (CO-2,
RM-2, SL-1, SL-2, SL-3, and KR-1) are probably the most
ecologically similar to the Douglas-fir study sites reported in
Table 1 and the foliage litterfall rates (0.12–0.15 kgm–2year–1)
are comparable with the minimum reported rates for Pacific
Northwest Douglas-fir stands (0.17–0.50 kgm–2year–1).
Fine woody fuel litterfall rates measured in this study for
those plots (0.001–0.139 kgm–2year–1) also compare well
with the Douglas-fir sites (0.005–0.129 kgm–2year–1). Foliar litterfall rates of the lodgepole sites (TF, RR-3, RR-4,
and CO-3 with rates of 0.12–0.15 kgm–2year–1) are about
one-half of those reported for lodgepole sites in Table 1
(0.362 kgm–2year–1), whereas Table 1 subalpine fir sites
have about double the rates (0.20–0.23 kgm–2year–1) of
the subalpine fir sites in this study (CO-3, SB-4, RM-4,
and KR-4; 0.036–0.157 kgm–2year–1). Large woody fuel
(logs) rates are highly variable in this study (0.0001–
0.207 kgm–2year–1), but they also seem to agree with those
reported for all studies in Table 1 (0.02–0.30 kgm–2year–1).
#
2008 NRC Canada
Keane
1439
Table 4. Results of the correlation analysis (coefficient r) of surface fuel litterfall rates
(kgm–2year–1) to both vegetation and environmental gradients.
Variable
Vegetation variables
FUELLOAD
BAREA
SAPAREA
SNAGBA
TPH
SAPPH
SPH
DOMDBH
AVEDBH
SNAGDBH
DOMHT
AVEHT
AGE
MAXAGE
LAI
Environmental variables
SOILDEPTH
SAND
SILT
CLAY
TMAX
TMIN
TDAY
PRCP
VPD
SRAD
NPP
OUTFL
Foliage
Twigs
Branches
Large
branches
Total
0.227
0.845
0.558
0.653
0.224
–0.358
0.269
0.491
0.359
0.376
0.646
0.533
0.153
0.065
0.722
0.116
0.777
0.217
0.497
0.392
–0.187
0.372
0.280
0.134
0.105
0.471
0.292
–0.003
0.000
0.610
–0.032
0.555
0.316
0.467
0.075
–0.290
0.383
0.385
0.482
0.086
0.531
0.264
–0.094
–0.029
0.543
–0.035
0.371
0.353
0.464
0.060
–0.100
0.279
0.137
0.072
0.301
0.291
0.170
0.029
–0.056
0.421
0.067
0.634
0.363
0.651
0.125
–0.321
0.504
0.271
0.360
0.255
0.475
0.445
0.031
–0.068
0.671
–0.333
–0.014
–0.176
0.239
0.340
0.452
0.383
–0.044
0.278
–0.173
–0.269
0.343
–0.048
–0.359
0.192
0.421
0.082
0.198
0.116
0.136
–0.009
–0.118
–0.242
0.384
–0.007
–0.159
0.092
0.178
0.310
0.263
0.310
0.053
0.262
–0.044
–0.045
0.251
–0.079
–0.161
0.162
0.097
0.094
0.374
0.171
0.319
–0.009
–0.503
0.055
0.024
–0.232
–0.113
–0.017
0.228
0.258
0.358
0.294
–0.009
0.195
–0.160
–0.161
0.103
Note: Variables are defined in Table 3. Correlation coefficients >|0.40| are given in boldface.
There are some limitations and shortcomings in this study
that might influence the litterfall findings. Several times, it
was impossible to empty litter traps on high-elevation plots
in the autumn because of early snowfalls, so there may have
been some decomposition losses because the summer’s litter
sat in the traps under the snow through the winter. Additionally, large snow banks on access roads sometimes delayed
visits in the spring for weeks, allowing the litter to sit in
traps under warm and moist conditions that were ideal for
decomposition. Many conifer tree species shed their foliage
during the late fall and early winter after the last trap visit
so many of the fallen needles remained in the snow above
the traps contributing to additional decomposition and wind
losses. Several traps were vandalized during the summers
causing gaps in the collection record for some plots. One
ponderosa pine plot (KR-2) experienced an autumn prescribed fire that burned all but one of the traps. The sorting
of foliage from other canopy material was a difficult and tedious task and was probably inconsistently done across the
12 field technicians involved in the project over the 10+
years of the study.
The lack of strong correlations of litterfall and decomposition rates with vegetation or environmental variables (Tables 4–6) may reflect missing biophysical variables, low
climate data quality and resolution, and the low number of
plots in this study. Important biophysical variables not included in this study are (i) morphological differences between tree species across the diverse sites, (ii) influence of
disturbance history on litterfall rates, and (iii) ecophysiological controls on environmental conditions. For example, leaf
longevity measurements differed across the species encountered on our plots ranging from 1–7 years on low-elevation
sites to 1–14 years on high-elevation sites because of differences in species composition (Keane et al. 2002). Litterfall
rates for all our plots were within a relatively small range
(0.057–0.136 kgm–2year–1) indicating that the simulated
and measured biophysical variables may have not had sufficient spatial and temporal resolution to consistently detect
the subtle differences in litterfall rates across the sites in
this study. In general, it appeared that litterfall was better
correlated with vegetation characteristics, whereas decomposition seemed to be related to environmental variables (Tables 4 and 5). The low number of plots in each vegetation
type by biophysical setting precluded a detailed investigation into the effect of species differences on fuel dynamics.
The number and size of litter traps used on the plot appears adequate for estimating litterfall for all components
but the large woody fuel size classes (logs and large
branches) (see Keane 2008). Results from an analysis of
temporal and spatial variance using bootstrap methods show
#
2008 NRC Canada
1440
Can. J. For. Res. Vol. 38, 2008
Fig. 5. Scatterplots of foliage litterfall and decomposition rates for the 28 plots in the study plotted with the vegetation and environmental
variables with the highest correlation: (a) and (b) leaf area index (LAI), (c) and (d) basal area, (e) and (f) dominant tree height, (g) and (h)
soil depth, and (i) and (j) minimum temperature.
#
2008 NRC Canada
Keane
1441
Table 5. Results of the correlation analysis (coefficient r) of surface fuel decomposition
rates (k values, year–1) to both vegetation and environmental gradients.
Variable
Vegetation variables
FUELLOAD
BAREA
SAPAREA
SNAGBA
TPH
SAPPH
SPH
DOMDBH
AVEDBH
SNAGDBH
DOMHT
AVEHT
AGE
MAXAGE
LAI
Environmental variables
SOILDEPTH
SAND
SILT
CLAY
TMAX
TMIN
TDAY
PRCP
VPD
SRAD
NPP
OUTFL
Foliage
Twigs
Branches
Large branches
–0.230
–0.248
–0.146
–0.015
0.055
0.015
0.004
–0.388
–0.263
–0.287
–0.398
–0.164
0.121
0.024
–0.243
–0.269
–0.392
–0.450
–0.231
0.020
0.259
–0.037
–0.505
–0.614
–0.564
–0.424
–0.388
–0.102
0.021
–0.579
0.431
–0.009
–0.021
0.045
0.020
–0.244
–0.065
–0.291
–0.049
–0.234
–0.081
0.136
0.160
0.099
0.017
0.359
–0.055
0.573
0.162
–0.339
–0.218
–0.203
0.144
0.443
0.025
0.042
0.187
0.393
0.177
–0.235
–0.219
–0.060
–0.106
0.276
–0.496
–0.458
–0.493
0.052
–0.473
–0.109
–0.275
–0.222
0.064
–0.256
0.303
0.085
–0.549
–0.391
–0.518
0.297
–0.593
–0.110
–0.007
–0.207
–0.213
–0.171
0.141
0.147
–0.031
–0.049
–0.036
0.372
–0.043
–0.198
0.216
0.257
–0.101
0.330
–0.218
–0.369
0.448
0.468
0.458
–0.111
0.441
–0.069
0.438
–0.078
Note: See Table 3 for variable abbreviations. Correlation coefficients >|0.40| are given in boldface.
that, although the collection design was sufficient (seven
traps capturing 90% variance) for fallen foliage and fine
woody fuels, the larger woody material was inadequately
sampled, because the traps were too few and too small to
capture the phenomenon of tree fall (Keane 2008). In retrospect, I probably should have sampled fallen logs and
branches across the entire plot and used the litter traps to
collect only the foliage and branchwood smaller than
25 mm, but this would have required extensive monitoring
of logs both inside and outside plot boundaries. Another alternative would have been to sample the logs across a longer
time span (>10 years), but this is problematic because staffing, funding, and transportation can be quite variable over
the ‡10 years of sampling. In an extension of this statistical
analysis (Keane 2008), it was determined that >30 plots of
7–9 traps each (>210 traps total) would be needed to
achieve a probability of detection >0.9 for logs for the entire
‡10 year record. Obviously, this large number of traps
would be quite costly and time consuming to install and
maintain. The tree life table and mortality rate approaches
used by other studies (e.g., Harmon and Hua 1991) appear
to be more effective, especially in ecosystems with large,
long-lived trees.
Fuel decomposition
The decomposition measurements of this study did not
match the rigor, detail, and scale of the litterfall measurements. Decomposition was only measured over a 3 year
time span, which was probably not long enough to adequately describe decay for the larger woody fuels. Decomposition was also only measured on five sites, and it did not
include logs and other canopy material. There were only
three sets of litterbags installed at each site, so a comprehensive analysis of variance such as that done for the litterfall
data was not possible with such a small sample. Moreover,
it was difficult to remove the material that had fallen on the
litterbags over the 3 years while they were in the field. Needles and small materials sometimes worked their way into
the bags through the coarse mesh, and decomposing material
below and on the top of the bags appeared to be brought
into the bag by soil macrofauna. Some bags were chewed
or torn apart by rodents and ungulates, and others were actually carried off site by unknown factors. These limitations
only affected around 16% of the samples.
Despite the limitations of the decomposition measurements, the measured rates seemed to compare quite well
with those measured in other studies (Table 1). The range
of k values for foliage decomposition measured in this study
for the Douglas-fir sites (0.085–0.205) are similar to those in
Table 1 (0.005–0.56). Similar results are found in the lodgepole sites (foliage: 0.104–0.247 in this study and 0.09–0.14
in Table 1) and subalpine fir sites (0.110–0.169 in this study
#
2008 NRC Canada
1442
Can. J. For. Res. Vol. 38, 2008
Table 6. Results of the regression analysis of surface fuel litterfall (kgm–2year–1) and decomposition rates (k, year–1)
to both the vegetation gradients and environmental gradients at the plot level using two variables (defined in Table 3)
at most.
Dependent variable
Accumulation rates (kgm–2)
Foliage
Twigs
Branches
Large branches
Total
Decomposition rates (k)
Foliage
Twigs
Branches
Large branches
Equation
df (error)
R2
y^
y^
y^
y^
y^
=
=
=
=
=
0.0173 + 0.00129 SNAGBA + 0.0019 BAREA
0.0144 + 0.00077 BAREA – 0.00037 SAND
exp(–6.962 + 0.0364 BAREA + 0.0358 AVEDBH)
exp(5.907 + 0.0034 SPH – 0.040 SRAD) – 0.001
exp(–2.668 + 0.0267 BAREA)
25
25
25
25
25
0.779
0.694
0.504
0.420
0.500
y^
y^
y^
y^
=
=
=
=
exp(–1.525 – 0.0076 DOMDBH)
0.146 – 0.0017 AVEDBH – 0.0023 LAI
exp(–3.312 + 0.017 FUELLOAD)
exp(–3.734 – 0.0077 MAXAGE + 0.0172 AGE)
18
13
13
14
0.167
0.571
0.166
0.370
Note: Significant correlation coefficients are given in boldface (p < 0.05).
and 0.09–0.17 in Table 1). This would indicate that the values calculated from this study should be useful for future
modeling efforts.
There are many reasons for the lack of strong correlation
between site factors and decomposition rates (Table 5), and
they are summarized in Prescott (2005). Decomposition is a
complex process that is highly influenced by local factors
acting at fine scales, such as tree spatial distribution, soil
type, and microclimate, and these fine scale influences and
associated variability may swamp effects of the coarse-scale
environmental variables used in this study to predict decay
rates (Kaarik 1974; Millar 1974; Moorhead and Sinsabaugh
2006). Moreover, the exponential function introduced by Olson (1963) does not seem to fit long-term trends of decomposition in many forests, so the k value may be
inappropriate for comparing decay rates across ecosystems.
I found that the exponential function used to determine k
often did not always fit the collected decomposition data
collected in this study because of the short time period
(3 years), slow rate of decomposition, and high variability
within a plot.
Conclusions
The correlation analysis shows that litterfall is related to
canopy characteristics, whereas decomposition is somewhat
related to biophysical site conditions. However, the low correlation coefficients suggest that there are many other environmental
variables
important
to
litterfall
and
decomposition, such as disturbance, species morphology,
and ecophysiological characteristics. The log woody fuel
deposition and decomposition estimates measured in this
study contain high error rates, but the fine fuel dynamics,
which is critically lacking in the literature, appear useful for
most models and management applications. The regression
analysis results provide a means to dynamically model the
measured fuel processes across landscapes using vegetation
and environmental variables. For example, the LAI spatial
product computed from MODIS imagery could be used to
map litterfall rates across the landscape using our regression
results (Table 6). These data can provide managers with valuable estimates of litterfall and decomposition rates that can
be used to determine the longevity of fuel treatments and
prioritize fuel treatment areas by calculating how long it
would take to accumulate enough surface fuels to ignite or
support a crown fire or kill overstory trees using the fire behavior models.
Acknowledgements
I acknowledge all those field technicians who spent
countless hours collecting, sorting, and weighing litter:
Todd Carlson, Kirsten Schmidt, Wayne Lynholm, Courtney
Couch, Laurie Dickinson, Myron Holland, Curtis Johnson,
Micha Krebs, Eric Apland, Daniel Covington, Amy Rollins,
and Ben McShan of the Rocky Mountain Research Station
Missoula Fire Sciences Laboratory. I also thank Joseph
White of Baylor University; Ceci McNicoll, USDA Forest
Service Gila National Forest; Wendel Hann, USDA Forest
Service Washington Office; Dan Fagre, USGS Glacier Field
Station; Dave Peterson, USDA Forest Service Pacific Northwest Research Station; Matt Rollins, Russell Parsons, Helen
Smith, Denny Simmerman, and Kathy Gray, USDA Forest
Service Rocky Mountain Research Station Missoula Fire
Sciences Laboratory; for field work, technical support, assistance in the analysis, and invaluable advice. Lastly, I thank
Roger Ottmar and Tom Spies, USDA Forest Service, Pacific
Northwest Research Station; Elizabeth Reinhardt, Rocky
Mountain Research Station; Jan van Wagtendonk, Yosemite
National Park; and four anonymous journal reviewers for insightful and helpful reviews. This work was partially funded
by the USGS National Biological Service and Glacier National Park’s Global Change Research Program under Interagency Agreements 1430-1-9007 and 1430-3-9005 and the
USGS CLIMET project.
References
Alexander, R.R. 1954. A comparison of growth and morality following cutting in old-growth mountain spruce stands. USDA
For. Serv. Rocky Mountain For. Range Exp Stn. Res. Note RN11.
Avery, C.C., Larson, F.R., and Schubert, G.H. 1976. Fifty-year records of virgin stand development in southwestern ponderosa
pine. USDA For. Serv. Gen. Tech. Rep. RM-22.
Berg, B., and Ekbohm, G. 1993. Decomposing needle litter in Pinus contorta (lodgepole pine) and Pinus sylvestris (Scots pine)
monocultural systems—is there a maximum mass loss? Scand.
J. For. Res. 8: 457–465.
#
2008 NRC Canada
Keane
Botkin, D.B. 1993. Forest dynamics: an ecological model. Oxford
University Press, New York.
Brais, S., Sadi, F., Bergeron, Y., and Grenier, Y. 2005. Coarse
woody debris dynamics in a post-fire jack pine chronosequence
and its relation with site productivity. For. Ecol. Manage. 220:
216–226. doi:10.1016/j.foreco.2005.08.013.
Bray, J.R., and Gorham, E. 1964. Litter production in forests of the
world. Adv. Ecol. Res. 2: 101–157.
Brown, J.K. 1970. A method for inventorying downed woody fuel.
USDA For. Serv. Gen. Tech. Rep. INT-16.
Busse, M.D. 1994. Downed bole-wood decomposition in lodgepole
pine forests of central Oregon. Soil Sci. Soc. Am. J. 58: 221–
227.
Carlton, D.W., and Pickford, S.G. 1982. Fuelbed changes with
aging of slash from ponderosa pine thinnings. J. For. 86: 91–
101.
Christiansen, E.C., and Pickford, S.G. 1991. Natural abatement of
fire hazard in Douglas-fir blowdown and thinning fuelbeds.
Northwest Sci. 65: 141–147.
Dimock, E.J. 1958. Litter fall in a young stand of Douglas-fir.
Northwest Sci. 32: 19–29.
Edmonds, R.L. 1979. Decomposition and nutrient release in Douglas-fir needle litter in relation to stand development. Can. J.
For. Res. 9: 132–140. doi:10.1139/x79-030.
Edmonds, R.L. 1987. Decomposition rates and nutrient dynamics in
small-diameter woody litter in four forest ecosystems in Washington, U.S.A. Can. J. For. Res. 17: 499–509. doi:10.1139/
x87-084.
Edmonds, R.L. 1991. Organic matter decomposition in western
United States forests. In Proceedings, Management and Productivity of Western-Montane Forest Soils, 10–12 Apr. 1990, Boise,
Idaho. Compiled by A.E. Harvey and L.E. Neuenschwander.
USDA For. Serv. Gen. Tech. Rep. INT-280. pp. 118–125.
Edmonds, R.L., Vogt, D.J., Sandberg, D.H., and Driver, C.H. 1986.
Decomposition of Douglas-fir and red alder wood in clear-cuttings. Can. J. For. Res. 16: 822–831. doi:10.1139/x86-145.
Edmonds, R.L., and Eglitis, A. 1989. The role of the Douglas-fir
beetle and wood borers in the decomposition of and nutrient release from Douglas-fir logs. Can. J. For. Res. 19: 853–859.
doi:10.1139/x89-130.
Fernandes, P.M., and Botelho, H.S. 2003. A review of prescribed
burning effectiveness in fire hazard reduction. Int. J. Wildland
Fire, 12: 117–128. doi:10.1071/WF02042.
Ferry, G.W., Clark, R.G., Montgomery, R.E., Mutch, R.W., Leenhouts, W.P., and Zimmerman, G.T. 1995. Altered fire regimes
within fire-adapted ecosystems. In Our living resources: a report
to the nation on the distribution, abundance, and health of U.S.
plants, animals, and ecosystems. U.S Department of the Interior,
National Biological Service, Washington, D.C. pp. 222–224.
Fogel, R., and Cromack, K., Jr. 1977. Effect of habitat and substrate quality on Douglas-fir litter decomposition in western
Oregon. Can. J. Bot. 55: 1632–1640. doi:10.1139/b77-190.
Fosberg, M.A. 1970. Drying rates of heartwood below fiber saturation. For. Sci. 16: 57–63.
US General Accounting Office (GAO). 2002. Severe wildland
fires: leadership and accountability needed to reduce risks to
communities and resources. Report to Congressional Requesters
GAO-02-259. US General Accounting Office, Washington, D.C.
Gottfried, G.L. 1978. Five-year growth and development in a virgin
Arizona mixed conifer stand. USDA For.Serv. Rocky Mountain
For. Range Exp. Stn. Res. Pap.RM-203.
Graham, R.L. 1982. Biomass dynamics of dead Douglas-fir and
western hemlock boles in mid-elevation forests of the Cascade
1443
Range. Ph.D. dissertation, Oregon State University, Corvallis,
Ore.
Grier, C.C., and Logan, R.S. 1977. Old-growth Pseudotsuga menziesii communities of a western Oregon watershed: biomass distribution and production budgets. Ecol. Monogr. 47: 373–400.
doi:10.2307/1942174.
Hagan, J.M., and Grove, S.L. 1999. Coarse woody debris: humans
and nature competing for trees. J. For. 97: 6–11.
Hann, W.J., Jensen, M.E., and Keane, R.E. 1988. Ecosystem management handbook. Ch. 4. ECODATA methods and field forms.
Northern Region Handbook, USDA Forest Service, Northern
Region, Missoula, Mont.
Harmon, M.E., and Hua, C. 1991. Coarse woody debris dynamics
in two old-growth ecosystems. Bioscience, 41: 604–610. doi:10.
2307/1311697.
Harmon, M.E., Franklin, J.F., Swanson, F.J., Sollins, P., Gregory,
S.V., Lattin, J.D., Anderson, N.H., Cline, S.P., Aumen, N.G.,
Sedell, J.R., Lienkaemper, G.W., Cromack, K., and Cummins,
K.W. 1986. Ecology of coarse woody debris in temperate ecosystems. Adv. Ecol. Res, 15: 133–302.
Hart, S.C., Firestone, M.K., and Paul, E.A. 1992. Decomposition
and nutrient dynamics of ponderosa pine needles in a Mediterranean-type climate. Can. J. For. Res. 22: 306–314. doi:10.1139/
x92-040.
Huebschmann, M.M., Lynch, T.B., and Wittwer, R.F. 1999. Needle
litterfall prediction models for even aged natural shortleaf pine
(Pinus echinata Mill.) stands. For. Ecol. Manage. 117: 179–
186. doi:10.1016/S0378-1127(98)00466-6.
Jenkins, M.A., Webster, R., Parker, G.R., and Spetich, M.A. 2004.
Coarse woody debris in managed central hardwood forests of Indiana, USA. For. Sci. 50: 781–792.
Johansson, M.-B. 1994. Decomposition rates of Scots pine needle
litter related to site properties, litter quality, and climate. Can. J.
For. Res. 24: 1771–1781. doi:10.1139/x94-229.
Kaarik, A.A. 1974. Decomposition of wood. In Biology of plant litter decomposition. Edited by C.H. Dickinson and G.J.F. Pugh.
Academic Press, London. pp. 129–174.
Keane, R.E. 2008. Surface fuel litterfall and decomposition in the
northern Rocky Mountains, USA. USDA For. Serv. Res. Pap.
RMRS-RP-70.
Keane, R.E., Ryan, K.C., and Running, S.W. 1996. Simulating effects of fire on northern Rocky Mountain landscapes with the
ecological process model Fire-BGC. Tree Physiol. 16: 319–331.
Keane, R.E., McNicoll, C., and Rollins, M.G. 2002. Integrating
ecosystem sampling, gradient modeling, remote sensing, and
ecosystem simulation to create spatially explicit landscape inventories. USDA For. Serv. Gen. Tech. Rep. RMRS-GTR-92.
Keenan, R.J., Prescott, C.E., Kimmins, J.P., Pastor, J., and Dewey,
B. 1996. Litter decomposition in western red cedar and western
hemlock forests on northern Vancouver Island, British Columbia. Can. J. Bot. 74: 1626–1634. doi:10.1139/b96-197.
Klemmedson, J.O. 1992. Decomposition and nutrient release from
mixtures of Gambel oak and ponderosa pine leaf litter. For.
Ecol. Manage. 47: 349–361. doi:10.1016/0378-1127(92)90284G.
Klemmedson, J.O., Meier, C.E., and Campbell, R.E. 1990. Litter
fall transfers of dry matter and nutrients in ponderosa pine
stands. Can. J. For. Res. 20: 1105–1115. doi:10.1139/x90-146.
Kueppers, L.M., Southon, J., Baer, P., and Harte, J. 2004. Dead
wood biomass and turnover time, measured by radiocarbon,
along a subalpine elevation gradient. Oecologia (Berl.), 141:
641–651. doi:10.1007/s00442-004-1689-x.
Laiho, R., and Prescott, C.E. 1999. The contribution of coarse
woody debris to carbon, nitrogen, and phosphorus cycles in
#
2008 NRC Canada
1444
three Rocky Mountain coniferous forests. Can. J. For. Res. 29:
1592–1603. doi:10.1139/cjfr-29-10-1592.
Laverty, L., and Williams, J. 2000. Protecting people and sustaining resources in fire-adapted ecosystems—a cohesive strategy.
Forest Service response to GAO Report GAO/RCED 99-65.
USDA Forest Service, Washington, D.C.
Mackensen, J., Bauhus, J., and Webber, E. 2003. Decomposition
rates of coarse woody debris—a review with particular emphasis
on Australian tree species. Aust. J. Bot. 51: 27–37. doi:10.1071/
BT02014.
Maguire, D.A. 1994. Branch mortality and potential litterfall from
Douglas-fir trees in stands of varying density. For. Ecol. Manage. 70: 41–53. doi:10.1016/0378-1127(94)90073-6.
Means, J.E., Cromack, K., Jr., and MacMillan, P.C. 1985. Comparison of decomposition models using wood density of Douglas-fir
logs. Can. J. For. Res. 15: 1092–1098. doi:10.1139/x85-178.
Meentemeyer, V. 1978. Macroclimate and lignin control of litter
decomposition rates. Ecology, 59: 465–472. doi:10.2307/
1936576.
Millar, C.S. 1974. Decomposion of coniferous leaf litter. In Biology of plant litter decomposition. Edited by C.H. Dickinson and
G.J.F.Pugh. Academic Press, London. pp. 105–129.
Moorhead, D.L., and Sinsabaugh, R.L. 2006. A theoretical model
of litter decay and microbial interaction. Ecol. Monogr. 76:
151–174. doi:10.1890/0012-9615(2006)076[0151:ATMOLD]2.0.
CO;2.
Nackaerts, K., Coppin, P., Muys, B., and Hermy, M. 2000. Sampling methodology for LAI measurements with LAI-2000 in
small forest stands. Agric. For. Meteorol. 101: 247–250. doi:10.
1016/S0168-1923(00)00090-3.
Olson, J.S. 1963. Energy storage and the balance of producers and
decomposers in ecological systems. Ecology, 44: 322–331.
doi:10.2307/1932179.
Pastor, J., and Post, W.M. 1985. Development of a linked forest
productivity – soil process model. Martin Marietta Energy Systems, Inc. for the US Department of Energy, Environmental
Sciences Division, Oak Ridge, Tenn. Environ. Sci. Div.
Publ. 2455.
Pfister, R.D., Kovalchik, B.L., and Arno, S. F., and Presby, R.C.
1977. Forest habitat types of Montana. USDA For. Serv. Gen.
Tech. Rep. INT-34.
Pearson, J.A., Knight, D.H., and Fahey, T.J. 1987. Biomass and nutrient accumulation during stand development in Wyoming lodgepole pine forests. Ecology, 68: 1966–1977. doi:10.2307/
1939887.
Prescott, C.E. 2005. Do rates of litter decomposition tell us anything we really need to know? For. Ecol. Manage. 220: 66–74.
doi:10.1016/j.foreco.2005.08.005.
Prescott, C.E., Taylor, B.R., Parsons, W.F.J., Durall, D.M., and
Parkinson, D. 1993. Nutrient release from decomposing litter in
Rocky Mountain coniferous forests: influence of nutrient availability. Can. J. For. Res. 23: 1576–1586. doi:10.1139/x93-198.
Prescott, C.E., Zabek, L.M., Staley, C.L., and Kabzems, R. 2000.
Decomposition of broadleaf and needle litter in forests of British
Columbia: influences of litter type, forest type, and litter mixtures. Can. J. For. Res. 30: 1742–1750. doi:10.1139/cjfr-30-111742.
Prescott, C.E., Hope, G.D., and Blevins, L.L. 2003. Effect of gap
size on litter decomposition and soil nitrate concentrations in a
high-elevation spruce–fir forest. Can. J. For. Res. 33: 2210–
2220. doi:10.1139/x03-152.
Reinhardt, E., and Keane, R.E. 1998. FOFEM—a first order fire effects model. Fire Manage. Notes, 58: 25–28.
Robertson, G.P., and Paul, E.A. 2000. Decomposition and soil or-
Can. J. For. Res. Vol. 38, 2008
ganic matter dynamics. In Methods in ecosystem science. Edited
by O.E. Sala, R.B. Jackson, H.A. Mooney, and R.W. Howarth.
Springer, New York. pp. 104–116.
Rollins, M.G., Keane, R.E., and Parsons, R.A. 2004. Mapping fuels
and fire regimes uisng remote sensing, ecosystem simulation,
and gradient modelling. Ecol. Appl. 14: 75–95. doi:10.1890/025145.
Rothermel, R.C. 1972. A mathematical model for predicting fire
spread in wildland fuels. USDA For. Serv. Pac. Intermountain
For. Range Exp. Stn. Res. Pap. INT-115.
Running, S.W., and Hunt, E.R., Jr. 1993. Generalization of a forest
ecosystem process model for other biomes, BIOME-BGC, and
an application for global-scale models. In Scaling physiological
processes: leaf to globe. Edited by J.R. Erhlinger and C.B. Field.
Academic Press, Inc, New York. pp. 141–157 .
Sanchez-Flores, E., and Yool, S.R. 2004. Site environment characterization of downed woody fuels in the Rincon Mountains, Arizona: regression tree approach. Int. J. Wildland Fire, 13: 467–
477. doi:10.1071/WF04015.
Sollins, P. 1982. Input and decay of coarse woody debris in coniferous stands in western Oregon and Washington. Can. J. For.
Res. 12: 18–28. doi:10.1139/x82-003.
Sollins, P., Cline, S.P., Verhoeven, T., Sachs, D., and Spycher, G.
1987. Patterns of log decay in old growth Douglas-fir forests.
Can. J. For. Res. 17: 1585–1595. doi:10.1139/x87-243.
Spies, T.A., Franklin, J.F., and Thomas, T.B. 1988. Coarse woody
debris in Douglas-fir forests of western Oregon and Washington.
Ecology, 69: 1689–1702. doi:10.2307/1941147.
Stohlgren, T.J. 1988. Litter dynamics in two Sierran mixed conifer
forests. II. Nutrient release from decomposing litter. Can. J. For.
Res. 18: 1136–1144. doi:10.1139/x88-175.
Stump, L.M., and Binkley, D. 1993. Relationships between litter
quality and nitrogen availability in Rocky Mountain forests.
Can. J. For. Res. 23: 492–502. doi:10.1139/x93-067.
Taylor, B.R., Prescott, C.E., Parsons, W.F.J., and Parkinson, D.
1991. Substrate control of litter decomposition in four Rocky
Mountain conifer forests. Can. J. Bot. 69: 2242–2250. doi:10.
1139/b91-281.
Thornton, P.E., Running, S.W., and White, M.A. 1997. Generating
surfaces of daily meteorological variables over large regions of
complex terrain. J. Hydrol. 190: 214–251. doi:10.1016/S00221694(96)03128-9.
Thornton, P.E., Law, B.E., Gholz, H.L., Clark, K.L., Falge, E.,
Ellsworth, D.S., Goldstein, A.H., Monson, R.K., Hollinger,
D.Y., Falk, M., Chen, J., and Sparks, J.P. 2002. Modeling and
measuring the effects of disturbance history and climate on carbon and water budgets in evergreen needleleaf forests. Agric.
For. Meteorol. 113: 185–222. doi:10.1016/S0168-1923(02)
00108-9.
Trofymow, J.A., Barclay, H.J., and McCullough, K.M. 1991. Annual rates and elemental concentrations of litter fall in thinned
and fertilized Douglas-fir. Can. J. For. Res. 21: 1601–1615.
doi:10.1139/x91-223.
Turner, J., and Long, J.N. 1975. Accumulation of organic matter in
a series of Douglas-fir stands. Can. J. For. Res. 5: 681–690.
doi:10.1139/x75-094.
Venables, W.N., and Ripley, B.D. 1999. Modern applied statistics
with SPLUS. Springer, New York.
Vogt, K.A., Grier, C.C., and Vogt, D.J. 1986. Production, turnover,
and nutrient dynamics of above- and belowground detritus of
world forests. Adv. Ecol. Res. 15: 303–377.
Webster, C.R., and Jenkins, M.A. 2005. Coarse woody debris dynamics in the southern Appalachians as affected by topographic
#
2008 NRC Canada
Keane
position and anthropogenic disturbance history. For. Ecol. Manage. 217: 319–330. doi:10.1016/j.foreco.2005.06.011.
White, D.E., Atzet, T., and Martinez, P.A. 2004. Relationship of
historic fire regimes to dead wood components in white fir forests of southwestern Oregon. Proc. Tall Timbers Fire Ecol.
Conf. 22: 117–124.
White, J.D., Running, S.W., Thornton, P.E., Keane, R.E., Ryan,
K.C., Fagre, D.B., and Key, C.H. 1998. Assessing simulated
ecosystem processes for climate variability research at Glacier
National Park, USA. Ecol. Appl. 8: 805–823. doi:10.1890/10510761(1998)008[0805:ASEPFC]2.0.CO;2.
White, M.A., Thornton, P.E., Running, S.W., and Nemani, R.R.
1445
2000. Parameterization and sensitivity analysis of the BIOMEBGC terrestrial ecosystem model: net primary production controls. Earth Interact. 4: 1–85. doi:10.1175/1087-3562(2000)
004<0003:PASAOT>2.0.CO;2.
Wright, K.H., and Lauterback, P.G. 1958. A 10-year study of mortality in a Douglas-fir sawtimber stand in Coos and Douglas
Counties, Oregon. USDA For. Serv. Pac. Northwest For. Range
Exp. Stn. Res. Pap. PNW-27.
Yavitt, J.B., and Fahey, T.J. 1982. Loss of mass and nutrient
changes of decaying woody roots in lodgepole pine forests,
southeastern Wyoming. Can. J. For. Res. 12: 745–752. doi:10.
1139/x82-113.
#
2008 NRC Canada